Efficient natural ventilation strategies could reduce 10 to 30 percent of a building’s overall energy consumption, but design procedures for optimal and robust natural ventilation systems are not well established. A major challenge to developing these procedures is the presence of several uncertainties in the models used for performance prediction. Deficiencies in design procedures and modeling can lead to suboptimal solutions from an energy-savings point of view, or to not identifying potential benefits and foregoing natural ventilation altogether.
This project’s goal is to address this challenge using an integrated system design and operational control strategy. By accounting for uncertainties during the design process we will enable the design of robust natural ventilation systems that can be operated using smart control systems during the building’s operational life. To achieve this goal we will: (1) develop a computational framework with uncertainty quantification to design flexible solutions for a variety of operating conditions, and (2) design a control system that will use real-time measurements in the operational building to learn which settings minimize energy consumption while maintaining ideal building conditions. This will optimize the energy savings of the system, and enable the more widespread use of natural ventilation systems in climates where natural ventilation is currently perceived as insufficiently robust or ineffective.
Research efforts will focus on the natural ventilation system in the Yang and Yamazaki Environment and Energy Building (Y2E2) building at Stanford University, which houses the Civil and Environmental Engineering department. The building’s hallways, open areas, and lounges are connected to 4 atria and cooled using a buoyancy-driven night flush, which replaces the warmer air inside the building with cooler outdoor nighttime air. The building’s HVAC system also consists of active heating or cooling within office spaces to maintain their temperatures. An extensive measurement system of 2400 sensors is in place and will be used for validating the model and as inputs to the control system. Developing a predictive model and operational control system that can account for and respond to uncertainties will help design natural ventilation systems that can be broadly implemented and achieve substantial energy use reductions.